This assumes that homosexuality is evenly distributed across age groups, something that is directly contradicted by the data. Your bias makes you think that this is caused by unacceptability of identification, but you would need to show that the change in acceptability has only affected younger generations.
As a counter-example, people who have worn LA lights sneakers are disproportionately under 30. It has become culturally acceptable to wear LA Lights sneakers over the last 15 years, perhaps older people are under-reporting their wearing of the sneakers?
> "How many Americans have ever cheated on a test?" and asking another set "Have you ever cheated on a test?" and then publishing an article about how "Americans Greatly Overestimate How Many Americans Have Cheated on a Test"
This metaphor seems directly contradicted by: > the change in the acceptability of identifying as LGBT over the last 70 years is pretty profound,
The change in ever cheating on a test over the last 70 years does not seem pretty profound to me.
It's clear that your personal bias is that there is an iceberg of non-identifying homosexuals who still feel oppressed. Even in SF, the number barely reaches 15% [1]
1: https://en.wikipedia.org/wiki/LGBT_demographics_of_the_Unite...